How to Monitor API Failures That Break Web Pages

Why API failures can look like page problems

Modern websites often use APIs for product data, pricing, carts, authentication, account information, search, inventory, personalization, and checkout. When one of those requests fails, the page may still be reachable while the actual experience breaks.

That makes API visibility important for website monitoring. Teams need to know whether the browser could reach the page and whether the page's own requests completed successfully — including all XHR and Fetch calls that run after the initial HTML loads.

How NorthDuty tracks API failures on web pages

NorthDuty intercepts both XHR and Fetch calls as Chromium renders each monitored page. Failed requests are grouped and classified as critical or non-critical before being reported and scored.

Critical API failures are first-party requests that directly serve your site's data or functionality. Each distinct group of critical failures reduces the Errors subscore by 25 points. Non-critical failures — analytics and tracking requests from domains like Google Analytics, Google Tag Manager, DoubleClick, Hotjar, Segment, Clarity, or Facebook — reduce the subscore by 10 points per group. The Errors component accounts for 20% of the overall 0–100 health score, with a floor of 10 regardless of how many failures are found.

Start by monitoring the web pages that depend on important first-party API calls: checkout, signup, login, dashboards, product pages, pricing, search, and account routes. For flows like checkout and signup, add user journey monitoring so the team knows whether an API failure blocks the customer action or only affects a secondary part of the page.

Conclusion

API failures can make a website unusable while the page itself still responds. Website monitoring needs browser-side API call visibility — not just URL availability — to catch these failures before customers do.

NorthDuty intercepts XHR and Fetch calls from a real Chromium browser, classifies each failure as critical or non-critical, and connects the result to the page health score and any journeys that depend on those requests. Critical API failures have a direct, measurable impact on the Errors subscore, which accounts for 20% of the total health score.